1
looris@Palace-of-the-Nine-Moons:/tmp $ bitcoind
looris@Palace-of-the-Nine-Moons:/tmp $ Bitcoin server starting
************************
EXCEPTION: St13runtime_error
CLevelDB(): error opening database environment Corruption: missing start of fragmented record(2)
bitcoin in AppInit()
This happens when I start bitcoind, any clue on what could I do to fix it?
I tried to run it with -rescan but the result was the same.
4
Just run ./bitcoind -reindex and this will fix your problem.
-rescan is for missing wallet transactions
0
Try to look in the ~/.bitcoin/db.log file for more information. It is likely that your only solution is to remove the block database (either blk0001.dat, blk0002.dat or blkindex.dat). DO NOT REMOVE YOUR wallet.dat file!!!
This means you will have to download the whole block chain again. This other question has tips on how to speed up that download.
Note that this may just be another case of the question asked here.
Don't advise people to wipe their data directories. Since 0.8, we have -reindex which rebuilds them without redownloading. – Pieter Wuille – 2013-10-09T17:51:55.917
I suspect it would be faster to redownload the blockchain than fix this, but here's where your error is being thrown from
– Nick ODell – 2013-02-13T23:25:40.2401@NickODell doing it, but it takes me more than 24h to download it all, if this bug happens again I'd really need a quicker way to fix it... – o0'. – 2013-02-13T23:42:53.523
1
Agreed. Actually, this seems more relevant to your problem. It would appear that if it loads a record of type kLastType which isn't preceeded by kFirstType, then it will put out that error message.
– Nick ODell – 2013-02-14T00:04:17.757@NickODell oh thank you, that would be useful. Now I'm re-downloading anyway, but I've not deleted the corrupted db yet... – o0'. – 2013-02-14T09:50:59.717
1You need -reindex, not -rescan (which is for finding missing wallet transactions). – Pieter Wuille – 2013-02-18T22:28:00.393